Association for Public Policy Analysis & Management

The effect of school leadership on student learning is second only to classroom teaching among school-level influences. Identifying those principal preparation programs whose graduates have a positive impact on student outcomes in the schools that they lead could provide educators and policymakers with valuable information with regard to designing and implementing more effective principal preparation programs and district partnerships, and meeting the goal of increasing student achievement.  While some prior studies have focused on the relative effectiveness of principals from a single program in a single district, ours is only the second large-scale study, following RAND’s New Leaders Evaluation, estimating the impact of principal preparation programs on student learning.  Our study is based on graduates from four principal preparation programs who served as principals in their program’s partner district between 2008-09 and 2014-15. These four programs were identified as having features advocated by leading principal preparation associations, and each program is located in a different region of the United States.  The analysis is based on student-level data for the 2008-09 through 2014-15 school years.  We evaluate the relative effectiveness of inexperienced principals from programs of interest using a comparative interrupted time series design.  We find little evidence that inexperienced principals from the programs of interest are more or less effective, on average, than inexperienced principals trained by other programs.  However, we do find significant variation in effectiveness among inexperienced principals.  In most cases the number of principals whose estimated contribution to student learning is significantly different than average is several times larger than one would expect to occur by chance, and the magnitudes of the estimated principal effects are consistent with previous research that uses student test scores to measure the relative performance of principals within schools.  We evaluate the relative effectiveness of both experienced and inexperienced principals from programs of interest using a within-student design to account for the possibility of nonrandom sorting of students into schools led by principals trained by one of the programs of interest.  Here we also find little evidence that principals from the programs of interest are more or less effective, on average, than principals trained by other programs.  While differences in principal effectiveness in these districts cannot be explained by the selected principal preparation programs, further exploration of the factors associated with variation in principal effectiveness may provide important insights on how best to develop excellent school principals, with the ultimate goal of increasing instructional effectiveness and student learning.